| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Agronomy, Iowa State Univ., Ames, IA 50011-1010
b Farm Services Agency, Elkader, IA
* Corresponding author (jsawyer{at}iastate.edu)
Received for publication November 6, 2006.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: adjR2, adjusted R2 index • CC, continuous corn • CM, chlorophyll meter • EONR, economic optimum nitrogen rate • ND, nitrogen rate difference from EONR • RCM, relative chlorophyll meter • SC, corn following soybean
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
For example, the Minolta SPAD 502 (Konica Minolta, Ramsey, NJ) CM is a handheld device that clamps on a leaf and measures light transmittance at 650 and 940 nm. The 650 nm wavelength coincides with the spectral region associated with maximum chlorophyll activity, while the 940 nm wavelength provides internal calibration to the instrument; compensating for leaf thickness, water status, and other plant factors. It has been well documented that CM readings are significantly related to the N status of corn plants (Piekielek and Fox, 1992; Schepers et al., 1992a; Wood et al., 1992; Blackmer and Schepers, 1994; Piekielek et al., 1995). This relationship is similar to N response with corn grain yield; is curvilinear and reaches a plateau at high leaf N concentrations (Wood et al., 1992; Blackmer and Schepers, 1994). This plateau indicates that CM readings are not sensitive to excess plant available N and luxury consumption. Readings produced by the CM have been found to be useful for determining plant N status when all factors other than N are constant. However, leaf greenness and therefore CM readings can be affected by multiple factors, such as drought stress, corn hybrid, stage of growth, and other nutrients. Chlorophyll meter readings, in field settings where all factors cannot be controlled, are most useful when compared with an adequately fertilized or non-N-limited reference within the same field. That is, creating a normalized value or sufficiency index (Schepers et al., 1992a, 1992b; Schepers, 1994). With the sufficiency index, when the value drops below 95% of the reference, N deficiency stress has occurred (Peterson et al., 1993; Blackmer and Schepers, 1995; Varvel et al., 1997). Fox et al. (2001) found that RCM values were 92% accurate in determining corn N status. The CM is easy to use, provides instantaneous results, and can be used to monitor the N status several times during the growing season.
Timing for determination of corn plant N status is important in relation to synchronization of soil N availability, N application, crop N demand, development of N stress, and N stress sensing. According to Dwyer et al. (1991), the narrow range of CM readings measured at growth stage V6 (Ritchie et al., 1993) makes it difficult to separate N-deficient from N-sufficient field areas. Varvel et al. (1997) found that only large N deficiencies could be detected using the CM at the V8 stage. However, Binder et al. (2000) measured significant grain yield reduction when N application was postponed beyond the V6 growth stage and little or no N was applied before sensing. Less than 20% of the total N uptake by corn occurs before V8 (Schepers et al., 1995). Russelle et al. (1983) points out that corn N uptake rate is affected by weather, planting date, and time of fertilizer application, but is generally greatest between V8 and R1 growth stages.
Nitrogen deficiencies detected late in the growing season (R4 to R5 growth stages) are more highly correlated with yield response to N than early season N stress detection (Blackmer and Schepers, 1995). Russelle et al. (1983) also found that when N applications were delayed to V16, the time of greatest N uptake was generally delayed until after R1. Scharf et al. (2002) found that when N application to corn was delayed until V12 to V16 there was a small but significant yield reduction. A greater reduction in yield resulted when application was delayed until R1; however, yield was still highly responsive to N.
To avoid substantial yield loss, in-season N should be applied before the VT to R1 growth stages (Russelle et al., 1983; Binder et al., 2000). Potential for yield reduction with late application is also related to the amount of soil N or fertilizer N available before determining N stress and making N applications. When using the CM approach for monitoring N status of corn, decisions on total crop N need should be more accurate when made later in the season and after significant N uptake. Although sensing N stress during reproductive growth stages provides a more accurate determination of crop N response and need, uptake of applied N and yield recovery that late in the season has not been successful. The window for sensing and in-season N application becomes narrow with mid-to-late vegetative growth stage timing, but it is this timing that researchers and producers are trying to make feasible for corn production.
Several researchers have estimated N application requirements using normalized CM readings. Blackmer and Schepers (1995) demonstrated that N fertigation soon after CM detection of a significant N stress (sufficiency index < 95%) could maintain an adequate crop N status and prevent yield loss. This technique requires continued monitoring of the corn crop with the CM and a small application each time N stress is detected. While this approach proved successful in irrigated corn production, regular monitoring is time intensive and the sensing may indicate N deficiency but not the total amount of fertilizer N needed to be applied for the entire season. In a regional project, Scharf et al. (2006) found that CM readings were related to EONR and corn yield response to N. The relationships were stronger for normalized than absolute readings, with CM readings taken later in the season, and when N had not been applied. The derived relationships should be useful for making N fertilization decisions. Francis and Piekielek (1999) provided an alternative management system for in-season N rate determination. This involved an accounting approach for several factors including plant N stress sensing, corn yield goals, manure credits, and stage of crop development. Usefulness of this multifactor approach, however, may be limited as N rates based on corn yield goals have not correlated well with optimum N rates (Doerge, 2002), and N credits from animal manure are difficult to estimate (Bausch and Diker, 2001). Also, plant N stress sensing was not used to directly determine N application rates.
There is a need for effective N rate determination in rain-dependent corn production where in-season N application would be limited to one application. The objective of this study was to investigate the potential for developing a calibration between corn plant N deficiency stress monitored with a CM and rate of fertilization required to provide economic optimal N.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The experimental design at all sites was a randomized complete block, with N rates replicated four times. The plot size was six or eight rows wide by 15 m in length. The first and second sets of trials had six N rates (0 to 225 kg N ha–1 in 45 kg increments). The third set of trials had seven N rates (0–270 kg N ha–1 in 45-kg increments), with one site having five N rates (0–270 kg N ha–1 in 67.5-kg increments). Nitrogen fertilizer was ammonium nitrate surface applied at crop emergence (46 sites), urea applied spring preplant and incorporated (26 sites), urea-ammonium nitrate solution applied spring preplant and incorporated (one site), or ammonium nitrate applied sidedress (one site). Locally adapted hybrids and corn production practices were used at the producer and research farm sites.
Corn plant N stress was determined with the Minolta SPAD 502 CM. Readings were taken following the procedure outlined in Peterson et al. (1993). Each CM reading was taken midway between the stalk and leaf tip, and midway between the midrib and leaf margin. Readings were taken from the uppermost leaf with the collar fully visible until tasseling, after which readings were taken from the ear leaf. Twenty to 30 plants were sensed in the middle rows of each plot. Chlorophyll meter readings were taken at the R1 corn growth stage (Ritchie et al., 1993) at all sites. At five of the third set of trial sites conducted in 2003 (the five sites with the same N rates), CM readings were taken at V8, V15, R1, and R3 corn growth stages in both SC and CC. Relative CM values were calculated for each site-year by dividing the average CM reading for each N rate by the corresponding average CM reading from the highest N rate. For sites with both crop rotations, RCM values were calculated within rotation.
Corn grain was hand harvested from the middle two rows (7.6-m length) of each plot after corn reached physiologic maturity, or harvested with a plot combine from the center rows of the entire plot length. Grain yields were adjusted to 155 g kg–1 moisture basis.
Statistical analyses were computed using SAS software (SAS Institute, 2003). Grain yield response to N rate for each site-year was analyzed by first using PROC GLM to determine whether N rate or mean N rate contrasted to zero N was significantly different (P 
0.10). The NLIN procedure was then used to fit regression models for those site-years identified as N responsive. The model that was statistically significant and with the highest coefficient of determination (R2) was selected. If the models had similar R2 values, the quadratic-plateau model was chosen. Response fit was also visually inspected against yield to confirm choice of appropriate model. The EONR for each site-year was determined from the regression model at a 0.10 N fertilizer to corn price ratio. If a site-year was nonresponsive, the EONR was set at zero. The ND was then calculated for each N rate within each site-year (EONR minus N rate).
To determine the relationship between ND and CM readings, and ND and RCM values, CM readings and RCM values were regressed against the corresponding ND for each rotation. PROC NLIN was used to investigate linear- and quadratic-plateau regression models. Models were deemed significant at P 0.05. Since CM readings were taken from multiple sites and years, PROC NLMIXED was used to allow inclusion of site-year as a random variable in the regression model. To assist with model comparisons, confidence limits (95%) for the regression parameters and the adjR2 value were calculated for each model.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
Using PROC NLMIXED to analyze CM readings by including site-year as a random variable resulted in an increased adjR2 value. For SC the index increased from 0.53 to 0.86, and for CC increased from 0.57 to 0.83. Analysis of RCM values by including the site-year variable resulted in only a small increase in adjR2 (for SC the index increased from 0.73 to 0.78, and for CC increased from 0.76 to 0.80). This indicates that site conditions add variability and are important factors influencing CM readings, and further reinforces the need to normalize readings to a non-N-limited reference if comparisons in N stress are to be made across varying field conditions, or between fields. Also, normalizing CM readings is an effective method to remove considerable variability due to site conditions. However, normalization did not remove all variability as there remained a scatter of RCM values, especially with deficit N (Fig. 2).
With SC, RCM values plateau at 0.99, and when the ND is zero (at optimal N) the RCM value is 0.97. When RCM values fall below 0.97, there is an ND deficiency with an increasingly smaller RCM value as the ND deficit increases. This 0.97 RCM value at optimal N is greater than the 0.95 value other researchers have found as a critical value or sufficiency index indicating plant N stress (Peterson et al., 1993). For N rates greater than zero ND (excess N), RCM levels off at 0.99 and does not increase with greater excess applied N. There are many RCM values at ND slightly less than optimum N (zero ND) that are close to the same values as when N is adequate or excess, that is, close to 0.97 RCM (Fig. 2). This is a portion of the N stress relationship that does not appear to differentiate well between adequate, slightly deficit, and excess N. Use of this RCM calibration for prediction of N applications could result in suggestions for low N rate application in-season when none is needed, or when there is only potential for a small yield increase to applied N. This could be partially alleviated by setting a minimum application rate.
We investigated fitting a linear-plateau model to the RCM values with SC. The adjR2 value was similar to the quadratic-plateau model (adjR2 = 0.72 and 0.73, respectively, for the linear- and quadratic-plateau), which is common to find with model fitting N response data. We believe that the quadratic-plateau model should be used to describe the relationship between RCM and ND because that model most often fits corn N response, the linear-plateau model resulted in the segmented regression line join point being at a deficit N rate (–22 kg N ha–1) with a plateau at 0.99 RCM, and the linear portion not reaching zero ND (intercept at 1.02 RCM). If the linear-plateau model were used to determine an N application rate, the first increment of suggested N based on RCM N stress sensing would occur at a high RCM value (0.99). That RCM value is larger than the often used 0.95 RCM to trigger an N application (Peterson et al., 1993). If the 0.95 RCM value were used instead, the ND rate would be –47 kg N ha–1, which is a high rate to correct a slight plant N stress. With the quadratic-plateau model, the intercept is at 0.97 RCM, which is closer to 0.95. Also, the ND rate at 0.95 RCM is –30 kg N ha–1, which would be a more moderate application with slight N deficiency.
Using a linear equation fit to RCM values from the V10 to R1 corn growth stages, Scharf et al. (2006) found the RCM value was 1.02 at zero EONR and the EONR was 42 kg N ha–1 at 0.95 RCM. Using a quadratic fit to RCM values in that study, the RCM value was 1.01 at zero EONR and the EONR was 44 kg N ha–1 at 0.95 RCM. These values are similar to those derived from the linear-plateau regression fit in our study, and larger than with the quadratic-plateau model. A plateau model fits our dataset due to inclusion of RCM values at excess N, which was not done by Scharf et al. (2006). Use of data with excess N is important to fully characterize N response and N stress sensing sensitivity around optimum N (slight deficit to excess N).
A quadratic-plateau regression model was fit for RCM values from the 28 CC site-years (Fig. 2). Despite being diverse datasets for the two rotations, the model parameters fall within the confidence limits for both CC and SC (Fig. 2), including the same adjR2 value (Table 1). This indicates that N stress development and the relationship between RCM values and ND is the same for corn in both rotations. Therefore, one calibration can be used for determining in-season N stress and application rates in SC and CC. While based on presensing applied N, expected in-season N need can be estimated with the regression model in Table 1. In the work by Scharf et al. (2006), SC and CC sites were successfully combined rather than separated in determination of the relationships between EONR, yield response to N, and sensing N stress.
At five of the third set of trial sites conducted in 2003, CM readings were taken at V8, V15, VT, R1, and R3 corn growth stages in both SC and CC. Chlorophyll meter readings, RCM values, and fitted regression models for each growth stage (CM and RCM vs. ND) are shown in Fig. 3 and 4 for the averages of the five sites. Model parameters are given in Table 2. As occurs with many factors that affect plant response to N, CM readings varied with growth stage. Readings across N stress levels were similar at the two early vegetative stages (V8 and V15) and at the two reproductive stages (R1 and R3) in both rotations (Fig. 3). This could be partially due to growth stage (and related N uptake) and to the change in leaf used for collecting readings between the vegetative and reproductive stages (upper fully expanded vs. ear leaf). When normalized to the highest N rate, the RCM values became more consistent between stages, especially at the V15 and R1 growth stages. This is similar to the analysis by Scharf et al. (2006) where RCM values from the V10 to R1 stages were combined into one group. Consistency in RCM values at V15 and R1 was greater in SC than CC. The RCM values at deficit N tended to be highest at V8, intermediate at V15 and R1, and lowest at R3. This indicates that as N stress development continued throughout the season, it was reflected in lower RCM values. Relative CM values at the V8 growth stage indicated N stress, but in SC not as much separation between deficit and excess N as at later growth stages.
|
|
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Since RCM values relate to the same ND with SC and CC, the same calibration for N application rate decisions based on RCM values can be used for corn in both rotations. The similarity in RCM values found at V15 and R1 growth stages suggest that there is a period of time during mid-to-late vegetative growth, rather than one critical time, to collect CM readings, obtain an indication of plant N stress and N rate to apply, and still provide time for in-season N application. The calibration between RCM values and ND found in this study could be used by producers to evaluate corn N stress in-season during mid-to-late vegetative growth stages across widely varying production conditions, determine if additional N is needed, and adjust N application rates.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. M. Samborski, N. Tremblay, and E. Fallon Strategies to Make Use of Plant Sensors-Based Diagnostic Information for Nitrogen Recommendations Agron. J., July 7, 2009; 101(4): 800 - 816. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Zotarelli, L. Avila, J. M. S. Scholberg, and B. J. R. Alves Benefits of Vetch and Rye Cover Crops to Sweet Corn under No-Tillage Agron. J., February 4, 2009; 101(2): 252 - 260. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Ruiz Diaz, J. A. Hawkins, J. E. Sawyer, and J. P. Lundvall Evaluation of In-Season Nitrogen Management Strategies for Corn Production Agron. J., November 7, 2008; 100(6): 1711 - 1719. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Ziadi, M. Brassard, G. Belanger, A. Claessens, N. Tremblay, A. N. Cambouris, M. C. Nolin, and L.-E. Parent Chlorophyll Measurements and Nitrogen Nutrition Index for the Evaluation of Corn Nitrogen Status Agron. J., August 11, 2008; 100(5): 1264 - 1273. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Solari, J. Shanahan, R. Ferguson, J. Schepers, and A. Gitelson Active Sensor Reflectance Measurements of Corn Nitrogen Status and Yield Potential Agron. J., May 7, 2008; 100(3): 571 - 579. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
